b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
The Brazilian Journal of
INFECTIOUS DISEASES www.elsevier.com/locate/bjid
Original article
Duplex realtime PCR method for Epstein–Barr virus and human DNA quantification: its application for post-transplant lymphoproliferative disorders detection María Dolores Fellner a,∗ , Karina Durand a , Marcelo Rodriguez b , Lucía Irazu b , Virginia Alonio a , María Alejandra Picconi a a
Oncogenic Viruses Service, Virology Department, National Institute of Infectious Diseases “Carlos G. Malbrán”, Av. Vélez Sársfield 563, C1282AFF Buenos Aires, Argentina b Operational Team Quality Management, Parasitology Department, National Institute of Infectious Diseases “Carlos G. Malbrán”, Av. Vélez Sársfield 563, C1282AFF Buenos Aires, Argentina
a r t i c l e
i n f o
a b s t r a c t
Article history:
Introduction: The quantification of circulating Epstein–Barr virus (EBV) DNA is used to moni-
Received 10 June 2013
tor transplant patients as an early marker of Post-Transplant Lymphoproliferative Disorders
Accepted 16 July 2013
(PTLD). So far no standardized methodology exists for such determination.
Available online 2 January 2014
Objective: Our purpose was to develop and validate a real-time PCR assay to quantify EBV DNA in clinical samples from transplant recipients.
Keywords:
Methods: A duplex real-time PCR method was developed to amplify DNA from EBV and
EBV
from a human gene. The EBV load was determined in peripheral blood mononuclear
real-time PCR
cells (PBMC), plasma and oropharyngeal tissue from 64 non-transplanted patients with
viral load
lymphoid-hypertrophy (Non-Tx), 47 transplant recipients without PTLD (Tx), 54 recipients
PTLD
with PTLD (Tx-PTLD), and 66 blood donors (BD). WinPEPI, version 11.14 software was used for statistical analysis. Results: Analytical validation: the intra and inter-assays variation coefficients were less than 4.5% (EBV-reaction) and 3% (glyceraldehyde 3-phosphate dehydrogenase – GAPDH reaction). Linear ranges comprised 107 –10 EBV genome equivalents (gEq) (EBV-reaction) and 500,000–32 human gEq (GAPDH-reaction). The detection limit was 2.9 EBV gEq (EBV-reaction). Both reactions showed specificity. Application to clinical samples: higher levels of EBV were found in oropharyngeal tissue from transplanted groups with and without PTLD, compared to NonTx (p < 0.05). The EBV load in PBMC from the groups of BD, Non-Tx, Tx and Tx-PTLD exhibited increasing levels (p < 0.05). In BD, PBMC and plasma, EBV loads were undetectable. Conclusions: The performance of the assay was suitable for the required clinical application. The assay may be useful to monitor EBV infection in transplant patients, in particular in laboratories from low-income regions that cannot afford to use commercial assays. © 2013 Elsevier Editora Ltda. All rights reserved.
∗
Corresponding author. E-mail addresses: [email protected], [email protected] (M.D. Fellner). 1413-8670/$ – see front matter © 2013 Elsevier Editora Ltda. All rights reserved. http://dx.doi.org/10.1016/j.bjid.2013.07.011
272
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
Introduction The quantification of Epstein–Barr virus (EBV) peripheral DNA is used to monitor transplant patients as an early marker of Post-Transplant Lymphoproliferative Disorders (PTLD).1–3 It has been proven that EBV load in peripheral blood samples from transplant patients with PTLD is higher than in transplant recipients without this disorder.4–6 Along the last decade, different quantitative PCR assays (semi-quantitative, competitive, real-time) have been used for surveillance, diagnosis, monitoring response to treatment, and determination of the degree of immunosupression to be applied. So far, there is no reference strategy to determine EBV load, including for instance the calibrator, the EBV region to be amplified, or the best sample type for identifying PTLD risk.7,8 No international standards were available until 2012, when the World Health Organization introduced the first WHO International Standard for Epstein–Barr virus, intended to be used for nucleic acid amplification techniques.9 Thus, the literature describes a variety of controls used to analytically validate EBV quantification assays, including cell lines,10–12 plasmids with EBV-genome fragments inserted,13,14 and commercially available controls containing viral particles.15,16 As no international standard or consensus-accepted control have been developed,17–19 each laboratory decided what calibrator to use to validate its own EBV quantification method. Different fragments of EBV genes were chosen for amplification in a variety of quantification assays, including repeated (BamHI-W region) or single (EBERs, EBNA-1, LMP-2, etc) viral genome regions20,21 with different degrees of sensitivity or accuracy, as previously described.22,23 Moreover, several blood sample types (peripheral blood mononuclear cells, plasma, and whole blood) were analyzed to identify PTLD. Most studies described and/or recommended using cell-associated blood samples over plasma/serum, but both specimen types appear to be informative and each laboratory determines its preference.24,25 Also, the extraction methods, the amount of sample to be analysed, the report format and the characteristics of the study populations vary between published data.7,20,21 Thus, all these factors have affected the comparison between methods. Several years ago, our laboratory developed a semiquantitative PCR strategy to measure EBV load26 and since then it has been used to monitor this viral infection in transplanted population from most of the institutions that perform organ transplantation in Argentina. The method is quite cumbersome and time-consuming; results demand at least 48 hours. Currently, real-time PCR quantification methods are widely applied to assess EBV load due to their advantages over conventional PCR assays.3,7,21 Moreover, the simultaneous amplification of an internal control along with the target DNA is widely used to detect the presence of inhibitors; it also allows to quantify the amount of sample present in the reaction, which permits viral load normalization.10,27 Despite the current availability of commercial assays, many laboratories from low income regions are unable to afford them; thus, less-costly in-house methods may be the only option to monitor EBV load in transplant patients. Their
development and validation could be extremely useful for the prevention of PTLD in these settings. Therefore, the aim of the present study was to develop and analytically validate a duplex real-time PCR assay to quantify EBV and human DNA in different types of clinical samples, in order to determine the EBV load in transplant patients regarding the risk of PTLD.
Materials and methods Patients and samples Children treated in “Prof. Dr Juan P. Garrahan” Pediatric Hospital, Austral University Hospital, “Sor María Ludovica” Children’s Hospital” and Favaloro Foundation, and blood ˜ donors of the “J. F. Muniz” Infectious Diseases Hospital were included as follows: (A) 64 non-transplanted patients with lymphoid hypertrophy in the oropharyngeal tissue. (B) 101 solid organ transplant patients (75 liver, 24 kidney, 2 heart), 54 of them with histological diagnosis of PTLD (including categories 1, 2, 3 and 4), according to the World Health Organization: IARC, 2008.28 (C) 66 blood donors, with negative results for all infections screened in routine blood bank protocol (hepatitis C virus, hepatitis B virus, human T lymphotropic virus I/II, human immunodeficiency virus, syphilis, brucellosis, and Chagas infection). All patients were infected with EBV according to the presence of IgG antibodies against viral capsid antigen (VCA) and/or viral DNA in peripheral blood. Patients in the transplant group were on an immunosuppressive regimen consisting of cyclosporine, tacrolimus or sirolimus, azathioprine or mycophenolate mofetil and steroids. Peripheral blood and oropharyngeal lymphoid tissue samples were taken due to oropharyngeal lymphoid hypertrophy, following the treatment protocol for transplant and nontransplanted patients. An informed consent was obtained in all cases as per the Helsinki declaration and other national and international regulations. Peripheral blood mononuclear cells (PBMC) and plasma were separated from 2.5 to 5 mL of EDTA-anticoagulated whole blood samples by centrifugation on a density gradient (Histopaque-1077, Sigma–Aldrich) and stored at −20 ◦ C. Oropharyngeal lymphoid tissue samples obtained through surgical removal from patients with lymphoid hypertrophy were stored at −80 ◦ C. Controls and Calibrators - EBV-specific reaction (EBNA-1 reaction) A plasmid containing a deleted fragment of the EBNA-1 coding gene from the EBV genome was used as calibrator of the real-time PCR quantification method. It had been previously developed in our laboratory to be used as competitor in
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
a quantitative competitive PCR. It was quantified by spectrophotometry at 260 nm and stored at −80◦ C. Calibrators were generated to exhibit the same characteristics as the clinical samples. Thus, for the EBNA-1 reaction, two types of calibrators were prepared:
273
PEBNA-1 was characterized by sequencing with the BigDye Terminator Sequencing kit v3.1, according to manufacturer’s recommendations in the Genetic Analyzer 3500 (Applied Biosystems-Hitachi). The sequence analysis was performed using the Sequencing Analysis Software v5.2 (Applied Biosystems). The PEBNA-1 sequence was compared with that of the prototypic EBV strain, B95-8, applying the BioEdit Sequence Alignment editor v7.0.9.
For the primers/probe design, the Primer Express software 2.0 (Applied Biosystems) was applied on the sequence of the PEBNA-1 to obtain MGB (Minor Groove Binding) type primers/probe, for the EBNA-1-reaction and on a conserved portion of the reference sequence of human GAPDH (AC 000144, GenBank) to obtain TAMRA (tetramethylrhodamine) type primers/probe for the GAPDH-reaction. In both cases, a pair of primers and probe for each reaction was selected, taking into account the most favourable condition with respect to their score and secondary structure; also, the possibility of non-specific cross-reactions was ruled out by analyzing their sequences with the Basic Local Alignment Search Tool (BLAST). The selected primers/probes were as follows: EBNA1 reaction: 5 CCGCTCCTACCTGCAATATCA 3 (forward primer) and 5 GGAAACCAGGGAGGCAAATC 3 (reverse primer); 5’ VICTGCAGCTTTGACGATGG-MGB 3’ (probe). They amplified a 73 base pair fragment. GAPDH reaction: 5 GGTGGTCTCCTCTGACTTCAACA 3 (forward primer); 5 GTGGTCGTTGAGGGCAATG 3 (reverse primer) and 5 FAM-CCACTCCT CCACCTTTGACGCTGG-3 TAMRA (probe). They amplified a 79 base pair fragment. Amplification was performed in a final reaction volume of 25 l, containing 1X TaqMan Universal Master Mix with AmpErase UNG (Applied Biosystems), 0.3 M of EBV-primers, 0.05 uM of GAPDH-primers, 0.1 M of EBV-probe and GAPDH-probe and the DNA to be amplified (calibrators for EBV or GAPDH reactions as described, 366 ng (equal to 105 cells) of DNA from PBMC or oropharyngeal tissue or a volume of DNA extracted from plasma (representing 30 l of plasma). The amplification was carried out using the 7500 real-time PCR System (Applied Biosystems) and the cycle conditions were as follows: 50 ◦ C 2 min; 95 ◦ C 10 min, followed by 45 cycles of 95 ◦ C 15 s, 60 ◦ C 1 min. The EBV load in PBMC DNA was expressed as the number of EBV genome equivalents per 105 PBMC. The normalized EBV load was estimated from the results of the EBNA-1 and GAPDH reactions (105 × EBNA-1 load/GAPDH load) and the unnormalized load from the result of the EBNA-1 reaction when measuring 366 ng of PBMC DNA estimated by spectrophotometry. The plasma EBV load was expressed as the number of EBV genome equivalents per mL of plasma multiplying the EBNA-1 reaction result by a factor of 33.3 (considering that the amount of plasma analyzed was 30 L, thus 30 L × 33.3 = 1000 L).
DNA extraction
Analytic validation
DNA from PBMC and oropharyngeal tissue samples was extracted as previously described.26 Plasma DNA was extracted using the QIAmp DNA mini kit (QIAgen) according to manufacturer’s instructions, taking into account the suggested recommendations for free viral DNA extraction.
Master batches of all controls and reagents were prepared for the analytical validation. The precision and dynamic range of the EBNA-1 and GAPDH reactions were determined in a one-day-run, with four replicates of each set of seven calibrators described over 20 consecutive days. The intra-assay (repeatability) and inter-assay (precision) variations were calculated using the CLSI/NCCLS. 2005. EP15-A2 procedure.29 The linear range was analysed using polynomial regression according to the CLSI/NCCLS. 2003. EP6-A procedure.30 The EBNA-1 reaction detection limit was determined using series of four samples prepared by diluting a high concentration EBV control (PEBNA-1 ) to dilutions containing 1, 2, 4 and 8
A Calibrators representative of cell-associated samples (peripheral blood mononuclear cells, tissues): serial dilutions of the PEBNA-1 were performed and placed on a background of 366 ng of commercially available human DNA (equivalent to 105 human cells).26 B Calibrators representative of cell-free samples (plasma/serum): serial dilutions of the PEBNA-1 were performed without the human DNA background. This allowed studying seven levels of EBV genome equivalents: 107 , 106 , 105 , 104 , 103 , 102 and 10 for each type of calibrator. - Human-specific reaction (GAPDH-reaction) Commercially available human DNA was used as a positive control for the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) reaction (Human Genomic DNA, Roche); which was quantified by spectrophotometry at 260 nm, aliquoted and stored at −80◦ C. Thus, seven calibrators were generated by 1/5 dilution of the commercially available human control, representing: 5 × 105 , 105 , 2 × 104 , 4 × 103 , 8 × 102 , 160 and 32 human cells. All these DNA concentrations were analyzed in the presence of high (106 EBV genome equivalents) or low (10 EBV genome equivalents) amounts, or in absence of the control PEBNA-1 .
PEBNA-1 sequencing
Real-time PCR method for EBV DNA quantification A duplex real-time PCR strategy that simultaneously amplifies portions of the EBV and of the human genome was applied, encoding the EBNA-1 protein and the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) enzyme respectively.
274
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
Table 1 – Description of the patients. Patients
Transplant patients
Non-transplanted patients Blood donors
a
N
Sex M/F
Agea
Agea at transplant
Organ transplanted
Cause of transplant
Non-PTLD
47
0.9
9.7 (0.5–19)
Liver (42) Kidney (5)
2.8 (0.5–12)
PTLD
54
1.5
8.0 (1–18)
Liver (35) Kidney (19) Heart (2)
4.7 (0.5–10)
64
1.8
–
–
–
66
2.3
7.0 (0.5–19) 36 (18–61)
–
–
Liver: FHF, BA EC, GL Kidney: CRF Liver: FHF, BA, EC, GL Kidney: CRF Heart: UC
Age in years, mean and range. FHF: fulminant hepatic failure; BA: biliary atresia; EC: sclerosing cholangitis; GL: glycogenosis type 1B and 3. CRF: Chronic renal failure. UC: unknown cause
copies of EBV genome equivalents, either in a background of human DNA representing 105 cells (366 ng) or without the DNA background. Twenty replicates of each sample were tested in three different runs. The detection limit was calculated using the probit regression function. The amplification efficiency of each run was calculated using the following formula: E = (10−1/m − 1) × 100.31 To analyze EBNA-1 reaction specificity, DNA from the Epstein–Barr virus (PEBNA-1 and RAJI cell line), from the human members of the Herpesviridae family (herpes simplex virus, cytomegalovirus, varicella-zoster virus, human herpesvirus 6 and human herpesvirus 8), from different origin human cells (fibroblast, human embryonic kidney (HEK-293) cell lines) and commercial source human DNA (Human Genomic DNA, Roche) were analysed. To analyze GAPDH reaction specificity, DNA from human cell lines (HEK-293 and human fibroblast), commercial source human DNA and plasmids containing different herpesvirus sequences (herpes simplex, varicella-zoster, cytomegalovirus, and human herpesvirus 6) were studied.
Statistical analysis The precision box-plot, dose response curve and regression analysis were performed with Microsoft Excel 2003. WinPEPI version 11.14 software was used for statistical analyses. Medians and confidence intervals were obtained with Describes, version 2.33. Differences between medians of continuous variables were analyzed using the Mann–Whitney or Kruskal–Wallis test for two groups (program Compare 2, version 2.57) or more than two groups (Etcetera, version 2.56). p-Values below 0.05 were considered significant.
10 CATTGAGTCG
B 95-8 PEBNA-1
20
40
70
B 95-8 PEBNA-1
GGCCCACTAA GGGAGTCCAT
B 95-8 PEBNA-1
TCATATATTT
110
80 TGTCTGTTAT
120
170
CAAAGCCCGC TCCTACCTGC
90 TTCATTGTCT
130
GCTGAGGGTT TGAAGGATGC
160
140
GATTAAGGAC
180
AATATCAAGG
190
GATGGAGTAG
220
ATTTGCCTCC
100 TTTTACAAAC
150 CTTGTTTTGC
200
TGACTGTGTG CAGCTTTGAC
Forward EBV-Primer
210 B 95-8 PEBNA-1
50
35 bp deletion 60
B 95-8 PEBNA-1
30
TCTCCCCTTT GGAATGGCCC CTGGACCCGG CCCACAACCT
EBV-MGB Probe
230
240
250
CCTATGGTGG AAGGGGCTGC
CTGGTTTCCA
Reverse EBV-primer
260 B 95-8 PEBNA-1
270
280
CGCGGAGGGT GATGACGGAG ATGACGGAGA
TGAAGGAG
Fig. 1 – Characterization of the EBV-fragment sequence inserted in the PEBNA-1 . The 35 bp deletion in the EBV fragment of PEBNA-1 and the forward/reverse primers. (– – –) and probe (—) used in the EBV-specific real-time PCR reaction are indicated.
B95-8 prototype strain, according to which both fragments were identical except for the 35 bp deletion.
Analytic validation of the real-time PCR EBV quantification assay - Precision
Results Description of the study population The population characteristics are described in Table 1.
Characterization of PEBNA.-1 control Fig. 1 shows the result of the analysis of the EBV fragment sequence inserted in PEBNA-1 compared with that of the
Fig. 2 shows the precision of the EBNA-1 reaction for cellassociated samples at the different EBV levels studied during the analytical validation. No differences were observed in the cycles’ threshold results obtained by the EBNA-1 reaction between the calibrators representing cell-associated samples or cell-free samples (data not shown). For the GAPDH reaction, the intra- and inter-run variation coefficients were less than 3%, at all concentrations studied (data not shown).
275
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
45 40
30 Probit
Cycle threshold
35
25 20 15 10 5
1,05 1 0,95 0,9 0,85 0,8 0,75 0,7 0,65 0,6 0,55 0,5 0,45 0,4 0,35 0,3 0,25 0,2 0,15 0,1 0,05 0 0
1
2
3
4
5
6
7
Log of the number of EBV genome equivalents
Fig. 2 – Precision box plot according to the EBV calibrators representative of cell-associated samples. The box plot shows the median, first and third quartiles of the cycle thresholds obtained for the EBV calibrators representative of cell-associated samples during the analytical validation; the minimum and maximum values are indicated (–).
- Linear range For the EBNA-1 reaction, the cycles thresholds’ regression analysis, obtained from the EBV calibrators, was applied to determine the linear range (Fig. 3). The standard error of the regression (taken as a measure of the model adjustment) was not higher than those for the second and third order models; it was concluded that the best match for this data set was the linear order model. Thereby, the reaction showed linearity across the whole range of concentrations studied (between 107 and 10 EBV genome equivalents), both for cell-associated and cell-free calibrators. A similar analysis for the GAPDH reaction showed a linear range between 500,000 and 32 human genome copies (data not shown). No differences were noted in the reaction results at all human DNA levels studied either in absence or presence of low or high concentrations of EBV genome equivalents (data not shown). - Detection limit Observed data
40,00
Linear model Quadratic model
35,00 Cycle threshold
1
2
3
4
5
Number of EBV gEq EBV genome equivalents
0
30,00
Fig. 4 – Detection limit of the EBV-specific reaction.
Fig. 4 shows the minimum concentration of EBV genome equivalents of the calibrator representing the cell-associated samples that could be readily detected. The probit regression analysis showed that the detection limit is about three EBV genome equivalents per 105 cells. A similar detection limit was obtained for calibrators representing cell-free samples. The GAPDH-reaction detection limit was not studied since the determination of the EBV load requires analysing close to 105 cells DNA (equivalent to 366 ng) from each sample. A very low amount of initial DNA may result from errors in loading the sample, DNA quality, or the presence of inhibitors; in all such cases the assay must be repeated. - Efficiency Both in the analytical validation and when studying the clinical samples, an amplification efficiency of 90–110% for both reactions was required in order to consider that an assay was acceptable. - Specificity When analysing the EBNA-1 reaction, a positive signal was only seen with the EBV controls’ DNA; no amplification was detected when analysing DNA from other human members of the Herpesviridae family (herpes simplex virus, cytomegalovirus, varicella-zoster virus, human herpesvirus 6 and human herpesvirus 8) or from human origin. Moreover, the GAPDH reaction gave a positive signal with different human DNAs but no signal with plasmid or various Herpesvirus (herpes simplex virus, cytomegalovirus, varicella-zoster virus and human herpesvirus 6) DNAs.
25,00
Correlation of normalized and unnormalized EBV loads measured in PBMC DNA
20,00
15,00 1
2
3
4
5
6
7
Log of the number of EBV genome equivalents
Fig. 3 – Dose response curve adjustment to linear and quadratic models.
The results of the EBV load measured in PBMC DNA representing 105 cells estimated by spectrophotometry (equal to 366 ng) (unnormalized load) and the GAPDH-reaction (normalized load) showed a linear correlation (data not shown). In some cases, when the amount of DNA was measured by
276
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
Table 2 – EBV viral load measure in different virus persistence sites. Study groups
Median of normalized EBV load (range) CIa
N Lymphoid tissue (EBV gEq/105 cells)
Blood donors
66
–
Non-transplanted patients with LH Transplant patients without PTLD Transplant patients with PTLD
64
4 (
The Brazilian Journal of
INFECTIOUS DISEASES www.elsevier.com/locate/bjid
Original article
Duplex realtime PCR method for Epstein–Barr virus and human DNA quantification: its application for post-transplant lymphoproliferative disorders detection María Dolores Fellner a,∗ , Karina Durand a , Marcelo Rodriguez b , Lucía Irazu b , Virginia Alonio a , María Alejandra Picconi a a
Oncogenic Viruses Service, Virology Department, National Institute of Infectious Diseases “Carlos G. Malbrán”, Av. Vélez Sársfield 563, C1282AFF Buenos Aires, Argentina b Operational Team Quality Management, Parasitology Department, National Institute of Infectious Diseases “Carlos G. Malbrán”, Av. Vélez Sársfield 563, C1282AFF Buenos Aires, Argentina
a r t i c l e
i n f o
a b s t r a c t
Article history:
Introduction: The quantification of circulating Epstein–Barr virus (EBV) DNA is used to moni-
Received 10 June 2013
tor transplant patients as an early marker of Post-Transplant Lymphoproliferative Disorders
Accepted 16 July 2013
(PTLD). So far no standardized methodology exists for such determination.
Available online 2 January 2014
Objective: Our purpose was to develop and validate a real-time PCR assay to quantify EBV DNA in clinical samples from transplant recipients.
Keywords:
Methods: A duplex real-time PCR method was developed to amplify DNA from EBV and
EBV
from a human gene. The EBV load was determined in peripheral blood mononuclear
real-time PCR
cells (PBMC), plasma and oropharyngeal tissue from 64 non-transplanted patients with
viral load
lymphoid-hypertrophy (Non-Tx), 47 transplant recipients without PTLD (Tx), 54 recipients
PTLD
with PTLD (Tx-PTLD), and 66 blood donors (BD). WinPEPI, version 11.14 software was used for statistical analysis. Results: Analytical validation: the intra and inter-assays variation coefficients were less than 4.5% (EBV-reaction) and 3% (glyceraldehyde 3-phosphate dehydrogenase – GAPDH reaction). Linear ranges comprised 107 –10 EBV genome equivalents (gEq) (EBV-reaction) and 500,000–32 human gEq (GAPDH-reaction). The detection limit was 2.9 EBV gEq (EBV-reaction). Both reactions showed specificity. Application to clinical samples: higher levels of EBV were found in oropharyngeal tissue from transplanted groups with and without PTLD, compared to NonTx (p < 0.05). The EBV load in PBMC from the groups of BD, Non-Tx, Tx and Tx-PTLD exhibited increasing levels (p < 0.05). In BD, PBMC and plasma, EBV loads were undetectable. Conclusions: The performance of the assay was suitable for the required clinical application. The assay may be useful to monitor EBV infection in transplant patients, in particular in laboratories from low-income regions that cannot afford to use commercial assays. © 2013 Elsevier Editora Ltda. All rights reserved.
∗
Corresponding author. E-mail addresses: [email protected], [email protected] (M.D. Fellner). 1413-8670/$ – see front matter © 2013 Elsevier Editora Ltda. All rights reserved. http://dx.doi.org/10.1016/j.bjid.2013.07.011
272
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
Introduction The quantification of Epstein–Barr virus (EBV) peripheral DNA is used to monitor transplant patients as an early marker of Post-Transplant Lymphoproliferative Disorders (PTLD).1–3 It has been proven that EBV load in peripheral blood samples from transplant patients with PTLD is higher than in transplant recipients without this disorder.4–6 Along the last decade, different quantitative PCR assays (semi-quantitative, competitive, real-time) have been used for surveillance, diagnosis, monitoring response to treatment, and determination of the degree of immunosupression to be applied. So far, there is no reference strategy to determine EBV load, including for instance the calibrator, the EBV region to be amplified, or the best sample type for identifying PTLD risk.7,8 No international standards were available until 2012, when the World Health Organization introduced the first WHO International Standard for Epstein–Barr virus, intended to be used for nucleic acid amplification techniques.9 Thus, the literature describes a variety of controls used to analytically validate EBV quantification assays, including cell lines,10–12 plasmids with EBV-genome fragments inserted,13,14 and commercially available controls containing viral particles.15,16 As no international standard or consensus-accepted control have been developed,17–19 each laboratory decided what calibrator to use to validate its own EBV quantification method. Different fragments of EBV genes were chosen for amplification in a variety of quantification assays, including repeated (BamHI-W region) or single (EBERs, EBNA-1, LMP-2, etc) viral genome regions20,21 with different degrees of sensitivity or accuracy, as previously described.22,23 Moreover, several blood sample types (peripheral blood mononuclear cells, plasma, and whole blood) were analyzed to identify PTLD. Most studies described and/or recommended using cell-associated blood samples over plasma/serum, but both specimen types appear to be informative and each laboratory determines its preference.24,25 Also, the extraction methods, the amount of sample to be analysed, the report format and the characteristics of the study populations vary between published data.7,20,21 Thus, all these factors have affected the comparison between methods. Several years ago, our laboratory developed a semiquantitative PCR strategy to measure EBV load26 and since then it has been used to monitor this viral infection in transplanted population from most of the institutions that perform organ transplantation in Argentina. The method is quite cumbersome and time-consuming; results demand at least 48 hours. Currently, real-time PCR quantification methods are widely applied to assess EBV load due to their advantages over conventional PCR assays.3,7,21 Moreover, the simultaneous amplification of an internal control along with the target DNA is widely used to detect the presence of inhibitors; it also allows to quantify the amount of sample present in the reaction, which permits viral load normalization.10,27 Despite the current availability of commercial assays, many laboratories from low income regions are unable to afford them; thus, less-costly in-house methods may be the only option to monitor EBV load in transplant patients. Their
development and validation could be extremely useful for the prevention of PTLD in these settings. Therefore, the aim of the present study was to develop and analytically validate a duplex real-time PCR assay to quantify EBV and human DNA in different types of clinical samples, in order to determine the EBV load in transplant patients regarding the risk of PTLD.
Materials and methods Patients and samples Children treated in “Prof. Dr Juan P. Garrahan” Pediatric Hospital, Austral University Hospital, “Sor María Ludovica” Children’s Hospital” and Favaloro Foundation, and blood ˜ donors of the “J. F. Muniz” Infectious Diseases Hospital were included as follows: (A) 64 non-transplanted patients with lymphoid hypertrophy in the oropharyngeal tissue. (B) 101 solid organ transplant patients (75 liver, 24 kidney, 2 heart), 54 of them with histological diagnosis of PTLD (including categories 1, 2, 3 and 4), according to the World Health Organization: IARC, 2008.28 (C) 66 blood donors, with negative results for all infections screened in routine blood bank protocol (hepatitis C virus, hepatitis B virus, human T lymphotropic virus I/II, human immunodeficiency virus, syphilis, brucellosis, and Chagas infection). All patients were infected with EBV according to the presence of IgG antibodies against viral capsid antigen (VCA) and/or viral DNA in peripheral blood. Patients in the transplant group were on an immunosuppressive regimen consisting of cyclosporine, tacrolimus or sirolimus, azathioprine or mycophenolate mofetil and steroids. Peripheral blood and oropharyngeal lymphoid tissue samples were taken due to oropharyngeal lymphoid hypertrophy, following the treatment protocol for transplant and nontransplanted patients. An informed consent was obtained in all cases as per the Helsinki declaration and other national and international regulations. Peripheral blood mononuclear cells (PBMC) and plasma were separated from 2.5 to 5 mL of EDTA-anticoagulated whole blood samples by centrifugation on a density gradient (Histopaque-1077, Sigma–Aldrich) and stored at −20 ◦ C. Oropharyngeal lymphoid tissue samples obtained through surgical removal from patients with lymphoid hypertrophy were stored at −80 ◦ C. Controls and Calibrators - EBV-specific reaction (EBNA-1 reaction) A plasmid containing a deleted fragment of the EBNA-1 coding gene from the EBV genome was used as calibrator of the real-time PCR quantification method. It had been previously developed in our laboratory to be used as competitor in
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
a quantitative competitive PCR. It was quantified by spectrophotometry at 260 nm and stored at −80◦ C. Calibrators were generated to exhibit the same characteristics as the clinical samples. Thus, for the EBNA-1 reaction, two types of calibrators were prepared:
273
PEBNA-1 was characterized by sequencing with the BigDye Terminator Sequencing kit v3.1, according to manufacturer’s recommendations in the Genetic Analyzer 3500 (Applied Biosystems-Hitachi). The sequence analysis was performed using the Sequencing Analysis Software v5.2 (Applied Biosystems). The PEBNA-1 sequence was compared with that of the prototypic EBV strain, B95-8, applying the BioEdit Sequence Alignment editor v7.0.9.
For the primers/probe design, the Primer Express software 2.0 (Applied Biosystems) was applied on the sequence of the PEBNA-1 to obtain MGB (Minor Groove Binding) type primers/probe, for the EBNA-1-reaction and on a conserved portion of the reference sequence of human GAPDH (AC 000144, GenBank) to obtain TAMRA (tetramethylrhodamine) type primers/probe for the GAPDH-reaction. In both cases, a pair of primers and probe for each reaction was selected, taking into account the most favourable condition with respect to their score and secondary structure; also, the possibility of non-specific cross-reactions was ruled out by analyzing their sequences with the Basic Local Alignment Search Tool (BLAST). The selected primers/probes were as follows: EBNA1 reaction: 5 CCGCTCCTACCTGCAATATCA 3 (forward primer) and 5 GGAAACCAGGGAGGCAAATC 3 (reverse primer); 5’ VICTGCAGCTTTGACGATGG-MGB 3’ (probe). They amplified a 73 base pair fragment. GAPDH reaction: 5 GGTGGTCTCCTCTGACTTCAACA 3 (forward primer); 5 GTGGTCGTTGAGGGCAATG 3 (reverse primer) and 5 FAM-CCACTCCT CCACCTTTGACGCTGG-3 TAMRA (probe). They amplified a 79 base pair fragment. Amplification was performed in a final reaction volume of 25 l, containing 1X TaqMan Universal Master Mix with AmpErase UNG (Applied Biosystems), 0.3 M of EBV-primers, 0.05 uM of GAPDH-primers, 0.1 M of EBV-probe and GAPDH-probe and the DNA to be amplified (calibrators for EBV or GAPDH reactions as described, 366 ng (equal to 105 cells) of DNA from PBMC or oropharyngeal tissue or a volume of DNA extracted from plasma (representing 30 l of plasma). The amplification was carried out using the 7500 real-time PCR System (Applied Biosystems) and the cycle conditions were as follows: 50 ◦ C 2 min; 95 ◦ C 10 min, followed by 45 cycles of 95 ◦ C 15 s, 60 ◦ C 1 min. The EBV load in PBMC DNA was expressed as the number of EBV genome equivalents per 105 PBMC. The normalized EBV load was estimated from the results of the EBNA-1 and GAPDH reactions (105 × EBNA-1 load/GAPDH load) and the unnormalized load from the result of the EBNA-1 reaction when measuring 366 ng of PBMC DNA estimated by spectrophotometry. The plasma EBV load was expressed as the number of EBV genome equivalents per mL of plasma multiplying the EBNA-1 reaction result by a factor of 33.3 (considering that the amount of plasma analyzed was 30 L, thus 30 L × 33.3 = 1000 L).
DNA extraction
Analytic validation
DNA from PBMC and oropharyngeal tissue samples was extracted as previously described.26 Plasma DNA was extracted using the QIAmp DNA mini kit (QIAgen) according to manufacturer’s instructions, taking into account the suggested recommendations for free viral DNA extraction.
Master batches of all controls and reagents were prepared for the analytical validation. The precision and dynamic range of the EBNA-1 and GAPDH reactions were determined in a one-day-run, with four replicates of each set of seven calibrators described over 20 consecutive days. The intra-assay (repeatability) and inter-assay (precision) variations were calculated using the CLSI/NCCLS. 2005. EP15-A2 procedure.29 The linear range was analysed using polynomial regression according to the CLSI/NCCLS. 2003. EP6-A procedure.30 The EBNA-1 reaction detection limit was determined using series of four samples prepared by diluting a high concentration EBV control (PEBNA-1 ) to dilutions containing 1, 2, 4 and 8
A Calibrators representative of cell-associated samples (peripheral blood mononuclear cells, tissues): serial dilutions of the PEBNA-1 were performed and placed on a background of 366 ng of commercially available human DNA (equivalent to 105 human cells).26 B Calibrators representative of cell-free samples (plasma/serum): serial dilutions of the PEBNA-1 were performed without the human DNA background. This allowed studying seven levels of EBV genome equivalents: 107 , 106 , 105 , 104 , 103 , 102 and 10 for each type of calibrator. - Human-specific reaction (GAPDH-reaction) Commercially available human DNA was used as a positive control for the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) reaction (Human Genomic DNA, Roche); which was quantified by spectrophotometry at 260 nm, aliquoted and stored at −80◦ C. Thus, seven calibrators were generated by 1/5 dilution of the commercially available human control, representing: 5 × 105 , 105 , 2 × 104 , 4 × 103 , 8 × 102 , 160 and 32 human cells. All these DNA concentrations were analyzed in the presence of high (106 EBV genome equivalents) or low (10 EBV genome equivalents) amounts, or in absence of the control PEBNA-1 .
PEBNA-1 sequencing
Real-time PCR method for EBV DNA quantification A duplex real-time PCR strategy that simultaneously amplifies portions of the EBV and of the human genome was applied, encoding the EBNA-1 protein and the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) enzyme respectively.
274
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
Table 1 – Description of the patients. Patients
Transplant patients
Non-transplanted patients Blood donors
a
N
Sex M/F
Agea
Agea at transplant
Organ transplanted
Cause of transplant
Non-PTLD
47
0.9
9.7 (0.5–19)
Liver (42) Kidney (5)
2.8 (0.5–12)
PTLD
54
1.5
8.0 (1–18)
Liver (35) Kidney (19) Heart (2)
4.7 (0.5–10)
64
1.8
–
–
–
66
2.3
7.0 (0.5–19) 36 (18–61)
–
–
Liver: FHF, BA EC, GL Kidney: CRF Liver: FHF, BA, EC, GL Kidney: CRF Heart: UC
Age in years, mean and range. FHF: fulminant hepatic failure; BA: biliary atresia; EC: sclerosing cholangitis; GL: glycogenosis type 1B and 3. CRF: Chronic renal failure. UC: unknown cause
copies of EBV genome equivalents, either in a background of human DNA representing 105 cells (366 ng) or without the DNA background. Twenty replicates of each sample were tested in three different runs. The detection limit was calculated using the probit regression function. The amplification efficiency of each run was calculated using the following formula: E = (10−1/m − 1) × 100.31 To analyze EBNA-1 reaction specificity, DNA from the Epstein–Barr virus (PEBNA-1 and RAJI cell line), from the human members of the Herpesviridae family (herpes simplex virus, cytomegalovirus, varicella-zoster virus, human herpesvirus 6 and human herpesvirus 8), from different origin human cells (fibroblast, human embryonic kidney (HEK-293) cell lines) and commercial source human DNA (Human Genomic DNA, Roche) were analysed. To analyze GAPDH reaction specificity, DNA from human cell lines (HEK-293 and human fibroblast), commercial source human DNA and plasmids containing different herpesvirus sequences (herpes simplex, varicella-zoster, cytomegalovirus, and human herpesvirus 6) were studied.
Statistical analysis The precision box-plot, dose response curve and regression analysis were performed with Microsoft Excel 2003. WinPEPI version 11.14 software was used for statistical analyses. Medians and confidence intervals were obtained with Describes, version 2.33. Differences between medians of continuous variables were analyzed using the Mann–Whitney or Kruskal–Wallis test for two groups (program Compare 2, version 2.57) or more than two groups (Etcetera, version 2.56). p-Values below 0.05 were considered significant.
10 CATTGAGTCG
B 95-8 PEBNA-1
20
40
70
B 95-8 PEBNA-1
GGCCCACTAA GGGAGTCCAT
B 95-8 PEBNA-1
TCATATATTT
110
80 TGTCTGTTAT
120
170
CAAAGCCCGC TCCTACCTGC
90 TTCATTGTCT
130
GCTGAGGGTT TGAAGGATGC
160
140
GATTAAGGAC
180
AATATCAAGG
190
GATGGAGTAG
220
ATTTGCCTCC
100 TTTTACAAAC
150 CTTGTTTTGC
200
TGACTGTGTG CAGCTTTGAC
Forward EBV-Primer
210 B 95-8 PEBNA-1
50
35 bp deletion 60
B 95-8 PEBNA-1
30
TCTCCCCTTT GGAATGGCCC CTGGACCCGG CCCACAACCT
EBV-MGB Probe
230
240
250
CCTATGGTGG AAGGGGCTGC
CTGGTTTCCA
Reverse EBV-primer
260 B 95-8 PEBNA-1
270
280
CGCGGAGGGT GATGACGGAG ATGACGGAGA
TGAAGGAG
Fig. 1 – Characterization of the EBV-fragment sequence inserted in the PEBNA-1 . The 35 bp deletion in the EBV fragment of PEBNA-1 and the forward/reverse primers. (– – –) and probe (—) used in the EBV-specific real-time PCR reaction are indicated.
B95-8 prototype strain, according to which both fragments were identical except for the 35 bp deletion.
Analytic validation of the real-time PCR EBV quantification assay - Precision
Results Description of the study population The population characteristics are described in Table 1.
Characterization of PEBNA.-1 control Fig. 1 shows the result of the analysis of the EBV fragment sequence inserted in PEBNA-1 compared with that of the
Fig. 2 shows the precision of the EBNA-1 reaction for cellassociated samples at the different EBV levels studied during the analytical validation. No differences were observed in the cycles’ threshold results obtained by the EBNA-1 reaction between the calibrators representing cell-associated samples or cell-free samples (data not shown). For the GAPDH reaction, the intra- and inter-run variation coefficients were less than 3%, at all concentrations studied (data not shown).
275
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
45 40
30 Probit
Cycle threshold
35
25 20 15 10 5
1,05 1 0,95 0,9 0,85 0,8 0,75 0,7 0,65 0,6 0,55 0,5 0,45 0,4 0,35 0,3 0,25 0,2 0,15 0,1 0,05 0 0
1
2
3
4
5
6
7
Log of the number of EBV genome equivalents
Fig. 2 – Precision box plot according to the EBV calibrators representative of cell-associated samples. The box plot shows the median, first and third quartiles of the cycle thresholds obtained for the EBV calibrators representative of cell-associated samples during the analytical validation; the minimum and maximum values are indicated (–).
- Linear range For the EBNA-1 reaction, the cycles thresholds’ regression analysis, obtained from the EBV calibrators, was applied to determine the linear range (Fig. 3). The standard error of the regression (taken as a measure of the model adjustment) was not higher than those for the second and third order models; it was concluded that the best match for this data set was the linear order model. Thereby, the reaction showed linearity across the whole range of concentrations studied (between 107 and 10 EBV genome equivalents), both for cell-associated and cell-free calibrators. A similar analysis for the GAPDH reaction showed a linear range between 500,000 and 32 human genome copies (data not shown). No differences were noted in the reaction results at all human DNA levels studied either in absence or presence of low or high concentrations of EBV genome equivalents (data not shown). - Detection limit Observed data
40,00
Linear model Quadratic model
35,00 Cycle threshold
1
2
3
4
5
Number of EBV gEq EBV genome equivalents
0
30,00
Fig. 4 – Detection limit of the EBV-specific reaction.
Fig. 4 shows the minimum concentration of EBV genome equivalents of the calibrator representing the cell-associated samples that could be readily detected. The probit regression analysis showed that the detection limit is about three EBV genome equivalents per 105 cells. A similar detection limit was obtained for calibrators representing cell-free samples. The GAPDH-reaction detection limit was not studied since the determination of the EBV load requires analysing close to 105 cells DNA (equivalent to 366 ng) from each sample. A very low amount of initial DNA may result from errors in loading the sample, DNA quality, or the presence of inhibitors; in all such cases the assay must be repeated. - Efficiency Both in the analytical validation and when studying the clinical samples, an amplification efficiency of 90–110% for both reactions was required in order to consider that an assay was acceptable. - Specificity When analysing the EBNA-1 reaction, a positive signal was only seen with the EBV controls’ DNA; no amplification was detected when analysing DNA from other human members of the Herpesviridae family (herpes simplex virus, cytomegalovirus, varicella-zoster virus, human herpesvirus 6 and human herpesvirus 8) or from human origin. Moreover, the GAPDH reaction gave a positive signal with different human DNAs but no signal with plasmid or various Herpesvirus (herpes simplex virus, cytomegalovirus, varicella-zoster virus and human herpesvirus 6) DNAs.
25,00
Correlation of normalized and unnormalized EBV loads measured in PBMC DNA
20,00
15,00 1
2
3
4
5
6
7
Log of the number of EBV genome equivalents
Fig. 3 – Dose response curve adjustment to linear and quadratic models.
The results of the EBV load measured in PBMC DNA representing 105 cells estimated by spectrophotometry (equal to 366 ng) (unnormalized load) and the GAPDH-reaction (normalized load) showed a linear correlation (data not shown). In some cases, when the amount of DNA was measured by
276
b r a z j i n f e c t d i s . 2 0 1 4;1 8(3):271–280
Table 2 – EBV viral load measure in different virus persistence sites. Study groups
Median of normalized EBV load (range) CIa
N Lymphoid tissue (EBV gEq/105 cells)
Blood donors
66
–
Non-transplanted patients with LH Transplant patients without PTLD Transplant patients with PTLD
64
4 (
